[mythtv] [mythtv-commits] Ticket #9566: Mythbackend reports "multiplex is not available" and afterwards the specified DVB tuner is busy for ever

Tomi Orava tomimo+mythtv-dev at ncircle.nullnet.fi
Fri Feb 18 10:50:04 UTC 2011


Hi,

> On 02/13/2011 02:14 AM, Tomi Orava wrote:
>> Does somebody have any ideas about the second(?) problem that my
>> system seems to have ie. it still has tendency to create recordings
>> which are missing the physical file ie. the file size is right from
>> the start marked as 'B' ?
>
> "file size...marked as 'B'" is because MythWeb shows you the file size
> using the most-appropriate units.  For a recording < 1024 bytes, the
> units would be bytes (B).  Since your recording has a size of 0 (because
> the file is non-existent), it shows the file size, but the 0 gets
> simplified out of the string, leaving just B.  All it's saying is it's a
> 0-byte recording.
>
> In general, any time mythbackend attempts to record a show and no data
> is written to disk /and/ no errors are shown in the mythbackend log, the
> problem is that the capture card or drivers provided absolutely no data
> for MythTV to write.  MythTV writes whatever data it's given, and if
> it's not given any, it doesn't write anything.  If the card/drivers
> don't tell MythTV there's an error, it doesn't write any errors to the log.
>
> I seriously think that your EIT (or something else) is stomping over
> your capture cards and puts one or more of them into an invalid state
> where it won't work properly (and doesn't give any errors) until reset
> (where a reset could require--depending on card/driver--one or more of
> restarting mythfrontend, reloading the card drivers, or rebooting the
> system).  Therefore, this is likely just a symptom of the EIT issue on
> the ticket, or something else (outside of MythTV) using your cards
> incorrectly, or a problem with the hardware or drivers.

Thanks for the clarification. I've been testing this system/problem  
since weekened in different ways and if I keep the EIT scanner  
functionality off on all but one card and also make sure that no  
encrypted channels have the "useonairguide" turned on, I haven't seen  
the previous "Multiplex is not available" errors or DVB—card hangs at  
all. The second (parent's 0.23 based system) works quite fine with  
these settings as well.

However, what comes to missing recordings (metadata exists, file  
doesn't, no errors of any kind) I don't believe that the problem is  
related to hardware or even about invalid installation or such.

1) The hardware has been working for 4 years now without big problems.
    For some reason I started seeing these problems myself on january
    after I did rescan the channels due to a channel numbering change.

2) My parents system is using exactly the same cards and gets its  
antenna signal
    from exactly the same source (physical distance is 30 meters) and that
    system seems to run pretty ok (I've seen occasional missing records during
    the past year ---> requires mythbackend restart, but mostly it works fine).

3) Most importantly this missing recordings problem seem to be  
directly related
    to the amount of virtual tuners per card. If I run with maximum 5 virtual
    tuners, I'll see the problem most propably in a 2-5 hours when trying to
    record everything available (load test for mythtv). However, if I lower the
    number of used virtual tuners to 3 per card, the stability increases quite
    significantly.

Do you have any pointers what is the actual file recording logic after  
a recording starts in tv_recp.cpp ? I'm interested in how we get from  
the dvbrecorder to ThreadedFileWriter as I still have some suspicions  
that some
signalling might get missed occasionally that might leave the file not  
being created at all. I have absolutely no proof of this at all, but  
I'm interested
in trying to figure out why I get these failures all the time the more virtual
tuners & real recordings I run.

Regards,
Tomi Orava






More information about the mythtv-dev mailing list